Module Code - Title:

CS6462 - PROBABILISTIC AND EXPLAINABLE AI

Year Last Offered:

2025/6

Hours Per Week:

Lecture

2

Lab

0

Tutorial

1

Other

3

Private

4

Credits

6

Grading Type:

N

Prerequisite Modules:

Rationale and Purpose of the Module:

This module commences by introducing probabilistic programming, an emerging area at the intersection of programming languages, probability theory, and probabilistic artificial intelligence. It then shows how to define probabilistic AI models and inference algorithms using executable code in probabilistic programming languages. Examples of modeling formalisms include generalized linear models, neural networks, Gaussian process, hierarchical Bayesian models , Bayesian networks and causal networks. Examples of inference approaches include Monte Carlo methods, expectation-maximization, and variational approaches. It focuses on explainability, or other procedural values in AI-enabled decisions. Emphasis is on applying these techniques to real data in a variety of application areas.

Syllabus:

1. Elements of probability calculus 2. Probabilistic programming 3. Bayesian Generalised Linear Models 4. Bayesian Neural Networks and Deep Neural Networks 5. Bayesian Nonparametric methods (Gaussian and Dirichlet processes) 6. Bayesian networks 7. Cognitive and casual models 8. Explainable AI

Learning Outcomes:

Cognitive (Knowledge, Understanding, Application, Analysis, Evaluation, Synthesis)

At the end of this module students will be able to: - Demonstrate an understanding of what a probabilistic program is - Write probabilistic models using a probabilistic programming language - Differentiate traditional machine learning from probabilistic machine learning - Model uncertainty - Know which algorithm to apply to a particular data set - Understand how to properly evaluate, explain and compare models

Affective (Attitudes and Values)

At the end of this module students will be able to: - demonstrate an understanding of the importance of explaining AI decisions - select the most appropriate approach to solve a given learning and decision problem

Psychomotor (Physical Skills)

N/A

How the Module will be Taught and what will be the Learning Experiences of the Students:

The module focuses on emerging trends in AI, probabilistic and explainable AI, and it will be delivered using a blended learning approach using traditional classroom lectures, on-line labs and tutorials.

Research Findings Incorporated in to the Syllabus (If Relevant):

Prime Texts:

Kevin Murphy (2013) Machine Learning: a probabilistic approach. , MIT press

Cameron Davidson-Pilon (2015) ayesian Methods for Hackers: Probabilistic Programming and Bayesian Inference , PEARSON

Judea Pearl, Madelyn Glymour, Nicholas P. Jewell (2016) Causal Inference in Statistics , Wiley

Other Relevant Texts:

Programme(s) in which this Module is Offered:

Semester(s) Module is Offered:

Spring

Module Leader:

emil.i.vassev@ul.ie